While much work has been devoted to studying transcriptional regulatory mechanisms, it has become increasingly clear that post-transcriptional processes also modulate the amount and utilization of RNA produced from a given gene. These post-transcriptional processes include nuclear post-transcriptional processes (e.g., splicing, polyadenylation, and transport) as well as cytoplasmic RNA degradation. All these processes contribute to the final steady-state level of a particular transcript. These points of regulation create a more flexible regulatory system than any one process could produce alone. For example, a short-lived message is less abundant than a stable one, even if it is highly transcribed and efficiently processed. The efficient rate of synthesis ensures that the message reaches the cytoplasm and is translated, but the rapid rate of degradation guarantees that the mRNA does not accumulate to too high a level. Many RNAs, for example the mRNAS for proto-oncogenes c-myc and c-fos, have been studied which exhibit this kind of regulation in that they are expressed at very low levels, decay rapidly and are modulated quickly and transiently under different conditions. See, M. Hentze, Biochim. Biophys. Acta 1090:281-292 (1991) for a review. The rate of degradation of many of these mRNAs has been shown to be a function of the presence of one or more instability/inhibitory sequences within the mRNA itself.
Some cellular genes which encode unstable or short-lived mRNAs have been shown to contain A and U-rich (AU-rich) INS within the 3' untranslated region (3' UTR) of the transcript mRNA. These cellular genes include the genes encoding granulocyte-monocyte colony stimulating factor (GM-CSF), whose AU-rich 3' UTR sequences (containing 8 copies of the sequence motif AUUUA) are more highly conserved between mice and humans than the protein encoding sequences themselves (93% versus 65%) (G. Shaw, and R. Kamen, Cell 46:659-667 (1986)) and the myc proto-oncogene (c-myc), whose untranslated regions are conserved throughout evolution (for example, 81% for man and mouse) (M. Cole and S. E. Mango, Enzyme 44:167-180 (1990)). Other unstable or short-lived mRNAs which have been shown to contain AU-rich sequences within the 3' UTR include interferons (alpha, beta and gamma IFNs); interleukins (IL1, IL2 and IL3); tumor necrosis factor (TNF); lymphotoxin (Lym); IgG1 induction factor (IgG IF); granulocyte colony stimulating factor (G-CSF), myb proto-oncogene (c-myb); and sis proto-oncogene (c-sis) (G. Shaw, and R. Kamen, Cell 46:659-667 (1986)). See also, R. Wisdom and W. Lee, Gen. & Devel. 5:232-243 (1991) (c-myc); A. Shyu et al., Gen. & Devel. 5:221-231 (1991) (c-fos); T. Wilson and R. Treisman, Nature 336:396-399 (1988) (c-fos); T. Jones and M. Cole, Mol. Cell Biol. 7:4513-4521 (1987) (c-myc); V. Kruys et al., Proc. Natl. Acad. Sci. U.S.A. 89:673-677 (1992) (TNF); D. Koeller et al., Proc. Natl. Acad. Sci. U.S.A. 88:7778-7782 (1991) (transferrin receptor (TfR) and c-fos); I. Laird-Offringa et al., Nucleic Acids Res. 19:2387-2394 (1991) (c-myc); D. Wreschner and G. Rechavi, Eur. J. Biochem. 172:333-340 (1988) (which contains a survey of genes and relative stabilities); Bunnell et al., Somatic Cell and Mol. Genet. 16:151-162 (1990) (galactosyltransferase-associated protein (GTA), which contains an AU-rich 3' UTR with regions that are 98% similar among humans, mice and rats); and Caput et al. Proc. Natl. Acad. Sci. 83:1670-1674 (1986) (TNF, which contains a 33 nt AU-rich sequence conserved in toto in the murine and human TNF mRNAs).
Some of these cellular genes which have been shown to contain INS within the 3' UTR of their mRNA have also been shown to contain INS within the coding region. See, e.g., R. Wisdom, and W. Lee, Gen. & Devel. 5:232-243 (1991) (c-myc); A. Shyu et al., Gen. & Devel. 5:221-231 (1991) (c-fos).
Like the cellular mRNAs, a number of HIV-1 mRNAs have also been shown to contain INS within the protein coding regions, which in some cases coincide with areas of high AU-content. For example, a 218 nucleotide region with high AU content (61.5%) present in the HIV-1 gag coding sequence and located at the 5' end of the gag gene has been implicated in the inhibition of gag expression. S. Schwartz et al., J. Virol. 66:150-159 (1992). Further experiments have indicated the presence of more than one INS in the gag-protease gene region of the viral genome (see below). Regions of high AU content have been found in the HIV-1 gag/pol and env INS regions. The AUUUA sequence is not present in the gag coding sequence, but it is present in many copies within gag/pol and env coding regions. S. Schwartz et al., J. Virol. 66:150-159 (1992). See also, e.g., M. Emerman, Cell 57:1155-1165 (1989) (env gene contains both 3' UTR and internal inhibitory/instability sequences); C. Rosen, Proc. Natl. Acad. Sci., U.S.A. 85:2071-2075 (1988) (env); M. Hadzopoulou-Cladaras et al., J. Virol. 63:1265-1274 (1989) (env); F. Maldarelli et al., J. Virol. 65:5732-5743 (1991) (gag/pol); A. Cochrane et al., J. Virol. 65:5303-5313 (1991) (pol). F. Maldarelli et al., supra, note that the direct analysis of the function of INS regions in the context of a replication-competent, full-length HIV-1 provirus is complicated by the fact that the intragenic INS are located in the coding sequences of virion structural proteins. They further note that changes in these intragenic INS sequences would in most cases affect protein sequences as well, which in turn could affect the replication of such mutants.
The INS regions are not necessarily AU-rich. For example, the c-fos coding region INS is structurally unrelated to the AU-rich 3' UTR INS (A. Shyu et al., Gen. & Devel. 5:221-231 (1991), and some parts of the env coding region, which appear to contain INS elements, are not AU-rich. Furthermore, some stable transcripts also carry the AUUUA motif in their 3' UTRs, implying either that this sequence alone is not sufficient to destabilize a transcript, or that these messages also contain a dominant stabilizing element (M. Cole and S. E. Mango, Enzyme 44:167-180 (1990)). Interestingly, elements unique to specific mRNAs have also been found which can stabilize a mRNA transcript. One example is the Rev responsive element, which in the presence of Rev protein promotes the transport, stability and utilization of a mRNA transcript (B. Felber et al., Proc. Natl. Acad. Sci. U.S.A. 86:1495-1499 (1989)).
It is not yet known whether the AU sequences themselves, and specifically the Shaw-Kamen sequence, AUUUA, act as part or all of the degradation signal. Nor is it clear whether this is the only mechanism employed for short-lived messages, or if there are different classes of RNAs, each with its own degradative system. See, M. Cole and S. E. Mango, Enzyme 44:167-180 (1990) for a review; see also, T. Jones and M. Cole, Mol. Cell. Biol. 7:4513-4521 (1987). Mutation of the only copy of the AUUUA sequence in the c-myc RNA INS region has no effect on RNA turnover, therefore the inhibitory sequence may be quite different from that of GM-CSF (M. Cole and S. E. Mango, Enzyme 44:167-180 (1990)), or else the mRNA instability may be due to the presence of additional INS regions within the mRNA.
Previous workers have made mutations in genes encoding AU-rich inhibitory/instability sequences within the 3' UTR of their transcript mRNAs. For example, G. Shaw and R. Kamen, Cell 46:659-667 (1986), introduced a 51 nucleotide AT-rich sequence from GM-CSF into the 3' UTR of the rabbit .beta.-globin gene. This insertion caused the otherwise stable .beta.-globin mRNA to become highly unstable in vivo, resulting in a dramatic decrease in expression of .beta.-globin as compared to the wild-type control. The introduction of another sequence of the same length, but with 14 G's and C's interspersed among the sequence, into the same site of the 3' UTR of the rabbit .beta.-globin gene resulted in accumulation levels which were similar to that of wild-type .beta.-globin mRNA. This control sequence did not contain the motif AUUUA, which occurs seven times in the AU-rich sequence. The results suggested that the presence of the AU-rich sequence in the .beta.-globin mRNA specifically confers instability.
A. Shyu et al., Gen. & Devel. 5:221-231 (1991), studied the AU-rich INS in the 3' UTR of c-fos by disrupting all three AUUUA pentanucleotides by single U-to-A point mutations to preserve the AU-richness of the element while altering its sequence. This change in the sequence of the 3' UTR INS dramatically inhibited the ability of the mutated 3' UTR to destabilize the .beta.-globin message when inserted into the 3' UTR of a .beta.-globin mRNA as compared to the wild-type INS. The c-fos protein-coding region INS (which is structurally unrelated to the 3' UTR INS) was studied by inserting it in-frame into the coding region of a .beta.-globin and observing the effect of deletions on the stability of the heterologous c-fos-.beta.-globin mRNA.
Previous workers have also made mutations in genes encoding inhibitory/instability sequences within the coding region of their transcript mRNAs. For example, P. Carter-Muenchau and R. Wolf, Proc. Natl. Acad. Sci., U.S.A., 86:1138-1142 (1989) demonstrated the presence of a negative control region that lies deep in the coding sequence of the E. coli 6-phosphogluconate dehydrogenase (gnd) gene. The boundaries of the element were defined by the cloning of a synthetic "internal complementary sequence" (ICS) and observing the effect of this internal complementary element on gene expression when placed at several sites within the gnd gene. The effect of single and double mutations introduced into the synthetic ICS element by site-directed mutagenesis on regulation of expression of a gnd-lacZ fusion gene correlated with the ability of the respective mRNAs to fold into secondary structures that sequester the ribosome binding site. Thus, the gnd gene's internal regulatory element appears to function as a cis-acting antisense RNA.
M. Lundigran et al., Proc. Natl. Acad. Sci. U.S.A. 88:1479-1483 (1991), conducted an experiment to identify sequences linked to btuB that are important for its proper expression and transcriptional regulation in which a DNA fragment carrying the region from -60 to +253 (the coding region starts at +241) was mutagenized and then fused in frame to lacZ. Expression of .beta.-galactosidase from variant plasmids containing a single base change were then analyzed. The mutations were all GC to AT transitions, as expected from the mutagenesis procedures used. Among other mutations, a single base substitution at +253 resulted in greatly increased expression of the btuB-lacZ gene fusion under both repressing and nonrepressing conditions.
R. Wisdom and W. Lee, Gen. & Devel. 5:232-243 (1991), conducted an experiment which showed that mRNA derived from a hybrid full length c-myc gene, which contains a mutation in the translation initiation codon from ATG to ATC, is relatively stable, implying that the c-myc coding region inhibitory sequence functions in a translation dependent manner.
R. Parker and A. Jacobson, Proc. Natl. Acad. Sci. U.S.A. 87:2780-2784 (1990) demonstrated that a region of 42 nucleotides found in the coding region of Saccharomyces cerevisiae MAT.alpha.1 mRNA, which normally confers low stability, can be experimentally inactivated by introduction of a translation stop codon immediately upstream of this 42 nucleotide segment. The experiments suggest that the decay of MAT.alpha.1 mRNA is promoted by the translocation of ribosomes through a specific region of the coding sequence. This 42 nucleotide segment has a high content (8 out of 14) of rare codons (where a rare codon is defined by its occurrence fewer than 13 times per 1000 yeast codons (citing S. Aota et al., Nucl. Acids. Res. 16:r315-r402 (1988))) that may induce slowing of translation elongation. The authors of the study, R. Parker and A. Jacobson, state that the concentration of rare codons in the sequences required for rapid decay, coupled with the prevalence of rare codons in unstable yeast mRNAs and the known ability of rare codons to induce translational pausing, suggests a model in which mRNA structural changes may be affected by the particular positioning of a paused ribosome. Another author stated that it would be revealing to find out whether (and how) a kinetic change in translation elongation could affect mRNA stability (M. Hentze, Bioch. Biophys. Acta 1090:281-292 (1991)). R. Parker and A. Jacobson, note, however, that the stable PGK1 mRNA can be altered to include up to 40% rare codons with, at most, a 3-fold effect on steady-state mRNA level and that this difference may actually be due to a change in transcription rates. Thus, these authors conclude, it seems unlikely that ribosome pausing per se is sufficient to promote rapid mRNA decay.
None of the aforementioned references describe or suggest the present invention of locating inhibitory/instability sequences within the coding region of an mRNA and modifying the gene encoding that mRNA to remove these inhibitory/instability sequences by making multiple nucleotide substitutions without altering the coding capacity of the gene.